In a perfect high-throughput screening world, hit-to-lead discovery scientists would have many chemical series from which to choose. Unfortunately, as Tara Stauffer, research scientist at Pharmacopeia (www.pharmacopeia.com), noted, “determining the optimal series for further chemical development can be a daunting task.”
At GTCbio’s “Assay Development & Screening Technologies Conference,” held in San Francisco last month, the evolution of the search process for hit-to-lead compounds was highlighted. A few of the presentations that featured new technologies designed to streamline the discovery process are reviewed in this article.
“A key concern in early-stage drug discovery is making sure that we work on physiologically relevant targets and assays,” said Berta Strulovici, Ph.D., vp, basic research at Merck & Co. (www.merck.com). Dr. Strulovici is also head of the automated biotechnology unit, a center of excellence for lead and target identification at Merck Research Laboratories.
“My department spends about 50 percent of its time working toward identifying novel compounds that affect molecular targets of interest by screening a large library of compounds against molecular targets. The other 50 percent is spent on the identification of proprietary targets with genome-scale RNA interference technologies.”
Dr. Strulovici presented specific examples of the benefits of implementing complex, multiparametric cellular assays on a large, industrial scale along the drug discovery pipeline—from small molecule lead ID, to target ID, to identifying potential responder populations.
For target ID using RNAi technology in human cell lines, multiparametric assays generate biological activity profiles that provide valuable insight into complex disease models. “Our infrastructure is unique in the industry,” explained Dr. Strulovici. “We are working with highly miniaturized, fully automated systems, that enable us to assess each gene in the genome as a potential target for drug discovery.”
Dr. Strulovici presented three case studies to demonstrate how this infrastructure impacts overall effectiveness in providing solid leads. The first study detailed mouse adult neuronal stem cell research to identify compounds that induce either proliferation or differentiation toward various cell lineages. The other two case studies focused on genome-scale siRNA screening, one for the identification of novel targets for Alzheimer disease, the other for the identification of responders in the clinic for an oncology program.
Combined with miniaturization technologies and sophisticated data-analysis tools, “we have identified novel targets for several therapeutic areas, which are now either at target-validation stage or have progressed to lead ID,” Dr. Strulovici said. “In addition to this, we are continually developing and refining technology to be able to perform the high-throughput biology needed to further research and drug discovery.
“As much as we work with artificial methods,” added Dr. Strulovici, “physiological assay systems are key. The easiest way is not the best way, since the physiological relevance may not be there. For the clinicians, you need to bear in mind that you do not work in isolation, the work you do needs to be relevant in the field. The team you build to get the job done is also key. Hire the best people possible and give them the freedom they need to make discovery possible.”
Highly engineered assays intended for high-throughput screening may not be ideal for selecting the best chemical series to pursue in a hit-to-lead campaign, according to Stauffer. “The bigger picture may include multiple assays to arrive at the best active series for chemical optimization,” she noted.
Pharmacopeia’s screening platform is based on its ECLiPS® (encoded combinatorial libraries on polymeric support) technology. Stauffer elaborated, “We basically go through three steps in the process: a primary HTS in which we screen the eluate from multiple beads; a follow up HTS, in which we screen single bead eluates from those sublibraries, which demonstrate activity in the assay; and finally, we submit beads for our unique decoding process, which is how we identify the structure of compounds in the active wells in step two.
“In the second step, we are searching for active compounds from a large random sample of beads out of a particular active sublibrary. Statistically, we will have multiple instances of the same bead (compound) in the screening plates. So in this process a truly active compound will produce multiple hits. A single compound might show up eight times, so the frequency of the hit demonstrates how strong the active signal is for that compound. This also reduces the incidence of false positives.”
The research she presented demonstrated how assays to explore differences in receptor binding, selectivity, and downstream functional activity were developed to reduce eight chemically distinct series, identified from a single HTS of six million compounds, to one pharmacologically relevant series that was suitable for further optimization. “Three were rejected out of hand for chemical properties, selectivity, and IP position,” noted Stauffer. “So the question remains, how do you choose the most viable option to pursue for development from the remaining five?”
The next round of compound vetting looked at binding and off-target testing; as a result, another compound series dropped out, she continued. “Looking at the functional pharmacology of the compounds was what eventually produced the best candidate series. Ultimately, evaluating compounds in the signaling pathway assay was what indicated which series would be the best to advance.”